Do Education Investment Areas make sense?

28th April 2022 by Timo Hannay [link]

Update 28th April 2022: See also this coverage from TES.

In summary (spoiler alert!), not really, at least statistically speaking. To understand why, read on.

Level best

The UK government's Levelling Up White Paper, published on 2nd February, declared its intention to support educationally underperforming parts of England:

"The UK Government will drive further school improvement in England through 55 new Education Investment Areas (EIAs) in places where educational attainment is currently weakest."

Among other things, multi-academy trusts will be encouraged to expand in these areas, some teachers will receive retention payments and there will be support for new 16-19 free schools (see our related analysis here). In addition, certain unspecified funds will be selectively channeled into EIAs.

On the face of it, this is sensible. English education is riven by geographical disparities: see any number of previous posts on this blog, particularly this overview of national trends, our survey of isolated schools and this analysis of coastal communities.

EIAs reflect many of these imbalances. Figure 1, below, shows the the mean values of various socioeconomic and educational indicators for EIA and non-EIA schools. EIA schools have higher proportions of pupils who are eligible for free school meals (FSM) and the Pupil Premium. They also have higher levels of local deprivation (as measured by the mean IDACI score within a 2km radius). Conversely, they show lower proportions of pupils reaching expected attainment levels by age 11, as well as lower Attainment 8 and Progress 8 scores at age 16.

(Use the menu to switch between indicators. Hover over the columns to see corresponding values and sample sizes.)

Re distributions

Given this, surely it makes sense to provide extra attention and resources to these underperforming parts of the country.

Yes and no. EIAs correspond to local authority (LA) areas. England contains 152 LAs, 55 of which have been designated as EIAs. On average, these underperform compared to the rest, so on average they need greater support. But that doesn't mean that every school or locality is doing badly. On the contrary, the variation in school performance within each LA is much greater than the variation between LAs.

Figure 2 shows the distributions of scores for the same metrics we saw in Figure 1. Measures of disadvantage such as FSM, Pupil Premium and IDACI show huge overlaps between EIA schools (red columns) and non-EIA schools (blue columns). It's a similar story when we look at measures of educational performance, such as Key Stage 2 attainment, Attainment 8 and Progress 8.

Fully 39% of non-EIA schools have FSM rates that are higher than the median EIA school; exactly the same goes for Pupil Premium rates. The corresponding proportion for IDACI is 34%. Looking at academic performance, 34% of non-EIA primary schools have lower KS2 attainment levels than the median EIA primary school. Among secondary schools, the corresponding proportions for Attainment 8 and Progress 8 are 38% and 39%, respectively. As a result, the EIA status of any individual school is a very poor predictor of its particular situation.

As a brief aside, it is worth dwelling on another of the government's levelling-up ambitions, which is to see 90% of all 11-year-olds achieving expected levels of attainment by 2030. As the Key Stage 2 attainment data here indicate, only a tiny proportion of schools – under 5% – achieved that target the last time primary pupils took SATs in 2019.

(Use the menu to switch between indicators. Hover over the columns to see corresponding values and sample sizes.)

The main lesson here is that EIAs contain many very strong schools, and there are plenty of underperforming schools in non-EIA parts of the country. To see this even more clearly, consider some specific examples:

• Leasowes High School, which is located in Dudley (an EIA) has 21% of its pupils on free school meals, while Four Dwellings Academy, located just over a mile away, over the border in Birmingham (not an EIA), has an FSM rate of 57%. Both have similar local deprivation rates and both have been rated 'Good' by Ofsted. On the face of things, it seems hard to argue that the former needs more support than the latter.
• Similarly, Sandwell Academy, which, appropriately enough, is located in Sandwell (an EIA) has 26% of pupils who are eligible for the Pupil Premium and is rated 'Outstanding', while Holyhead School, which is in Birmingham (again, not an EIA) has 60% of such pupils and is rated 'Good'. Both have broadly similar local deprivation rates, though Holyhead's is slightly higher. If we were going to pick one of these for greater support it would probably be Holyhead.
• On the academic front, St Joseph's College an 'Outstanding' school in Stoke-on-Trent (an EIA), achieved a very high average Attainment 8 score of 61 the last time GCSEs were sat in 2019. Meanwhile, Clayton Hall Academy less than a mile away, but in Staffordshire (not an EIA), is rated 'Requires Improvement' and obtained an Attainment 8 score of 47. This is perfectly respectable, but obviously much lower than its supposedly disadvantaged neighbour.
• Even more starkly, Levenshulme High School in Manchester (an EIA) is rated 'Outstanding' and has a Progress 8 score of 1.14 (which is to say that its GCSE pupils overperformed by more than one grade per subject), while Reddish Vale High School less than 2 miles away in Stockport (not at EIA) is rated 'Inadequate' and has a Progress 8 score of -0.56.
• Among primary schools, Pennyman Primary Academy, located in Middlesborough (an EIA), got 59% of its pupils to expected levels of attainment by age 11. This is a bit below average but certainly not terrible in the scheme of things, especially given its high proportion of poor families (59% of pupils are eligible for FSM). Consistent with this, academic progress, which takes into account pupils' starting points, is good. A few hundred metres down the road, Saint Gabriel's Catholic Primary School, which is in Redcar and Cleveland (not an EIA), has lower in-school deprivation but got only 30% of its 11-year-olds to expected levels of attainment. Both schools are rated 'Good' by Ofsted.

We could go on, but you get the idea. Educational disadvantage and underperformance often don't conform to simple stereotypes, and are certainly no respecters of administrative boundaries. Attention really needs to be directed at neighbourhoods and schools rather than whole LAs. We came to a similar conclusion in our analysis of coastal communities, which showed that, consistent with conventional wisdom, coastal locations do indeed tend to have higher levels of disadvantage, but also that this is very unevenly distributed, so to generalise about them is to over-simplify. We can say much the same about LAs and EIAs.

Of course, this doesn't mean that EIAs make no political sense: the levelling-up agenda is deeply rooted in voters' sense of place, and when it comes to schools this might be reasonably equated to LAs. It is merely to point out that EIAs are a blunt tool. England's inequalities, geographical or otherwise, ultimately need to be addressed at the level of individual neighbourhoods, schools and even pupils.